## Roundup 10/30/10

Football match-ups are intransitive. Team A beats team B, who beats team C. So we'd expect team A to beat team C, right? Not always. So far this year there have been 11 such cases of 3-team intransitivity.

"Second team" All-Pros from the AP are selected the wrong way.

How many Super Bowl-winning QBs are playing in the league in any given season?

Mitchel Lichtman on luck in baseball, but it applies to sports in general. It applies to the whole world, for that matter.

For those who missed it, Andy Steiner takes a look at 'momentum' following interceptions at the Community Site. Awesome.

LaDanian Tomlinson says the current Jets are better than any Chargers team. I'm not even sure if they're better than the current Chargers.

Does "playoff experience" matter, at least in baseball? It's something the sports media likes to tell us is important. They have to fill up those 2-hour long pre-game shows somehow. (Helmet-knock: Tango)

Football doesn't have playoff series the way baseball, basketball, or hockey does. But here's an interesting application that combines WPA with the concept of game leverage in a series, something I wrote about last spring. This is another media talking point--the "must win" game. (Another h/k to Tango)

I've been periodically updating the publicly-available play-by-play data for 2010, but John Candido has been going the extra mile and making it available each week.

Football Outsiders attempts to analyze the Patriots' 4th down and 1 decision to go for it last week. It's an excruciating 1,215 words, and it's chock full of assumptions and what the author calls "guesstimates." Ironically, the post is titled "Quick Reads." There is a danger in using game-specific intuitive guesses of probabilities. It's mostly a futile exercise of the small-sample fallacy, also known as the 'hasty-generalization'. Teams only have about 10 full drives a game, not nearly enough to know much about their true respective probabilities on any one drive, and certainly not on any one play. Even the best and worst NFL teams are surprisingly close in ability on any one play. It's only when the slight marginal differences of 120 plays accumulate do we (usually) see the better team come out on top. To be fair, game-specific considerations do matter, just not nearly as much as most people would think. And guessing at them based on a handful of previous drives induces more error than it helps.

The industrial organization of the Miami Heat. Hierarchy matters.

Carl Bialik on the Chargers, Cowboys and the inadequacy of total yards as a measure of a team.

The helmet-to-helmet issue consumed all the oxygen in the room last Sunday. So when teams collectively scored the second most points in 20 years last week, many pundits said it could be due to the threat of heavy fines and suspensions. But this probably isn't the case.

### 7 Responses to “Roundup 10/30/10”

1. SportsGuy says:

Where are this week's Adjusted S/R ratings, as displayed in last week's post here:

2. Anonymous says:

Why on earth would anyone take anything the AP votes on seriously? AP All-Pro selections, you have got to be kidding me? Next thing you are going to be telling me we actually select the most qualified person to run the country every four years instead of the one who is the best fundraiser/liar/navigator of his political party.

The AP is demonstrably certifiably, a completely negligent group of morons. Same with the academy, the electorate, or pretty much any group of people who are voters but not selected on the basis of merit.

3. Dan R says:

Any voting procedure has "issues". It's just a debate about how bad we feel about different results.

See the link for a technical description, but no voting system can satisfy the following three criteria (at the same time).

If everyone prefers A over B, so do the combined results.

If no one switches their preference between A and B, then the overall result between A and B doesn't change.

The vote actually matters (there's not a dictator).

The AP All Pro balloting method is a specific type of group preference generation, and it does not meet all three of the above.

4. James says:

In response to the inadequacy of total yards, I've wondered if the percentage of total potential yards gained would be a better statistic. Something a little more involved could be better, but I would think this would be a very simple yet informative statistic.

5. Ian Simcox says:

Wow that FO article is badly written. If anything, that's a lesson in how not to present stats. Even me, who's quite interested in the stats, struggled to keep up.

6. Andy says:

About the intransitive article; i would be really interested to see how many football match ups are trasitive, team A beats team B, team B beats team C, so then team A beats team C. I think it would be interesting to compare how many of those there are to the 11 mentioned in the article. This would be an interesting study to do over multiple seasons to test how....linear (i guess?)..NFL games are.

7. Brian Burke says:

Andy-Check out beatpaths.com